Coverity has updated its software application development testing platform with a new component named Test Advisor, which includes impact analysis technology to focus testing on high risk code. The firm defines "high risk" code to include "changed code" and "code impacted by a change". The goal here is an opportunity to alert developers of code not covered by unit tests.
The firm is politely critical of "traditional unit testing" as we might understand it, saying that while valuable, it has been inefficient, "failing to focus squarely on critical parts of the code" and also "lacking insight into change impact" in general.
Coverity says that it has extended its Static Analysis Verification Engine (SAVE) with "change impact analysis" and that this ensures code changes are sufficiently tested by mapping code and function dependencies to analyze impacted code related to a change.
NOTE: This function works for both code that has been directly modified and unmodified code that is impacted by the code change.
New Java web application analysis technology is also available including an enterprise framework analyser, which works to augment static source code analysis to identify new types of defects. The end result, Coverity hopes, is that false positives are ultimately minimized.
The firm also offers remediation tools to fix quality defects, security defects, test violations, and third-party analysis defects. Applicable to a unified workflow, this technology offers defect-specific guidance to help developers fix security defects "without specialized knowledge".
There is an open platform API here, so programmers can (potentially) integrate third-party software analysis results into the Coverity Development Testing Platform. Plus, the firm supports bidirectional integration with existing lifecycle tools including IDEs, source control management, test coverage, bug tracking, build and continuous integration, and application lifecycle management solutions to make development testing "a natural part" of the software development process.